random.pl -- Random numbers

This library is derived from the DEC10 library random. Later, the core random generator was moved to C. The current version uses the SWI-Prolog arithmetic functions to realise this library. These functions are based on the GMP library.

author: - R.A. O'Keefe, V.S. Costa, L. Damas, Jan Wielemaker
See also: - Built-in function random/1: A is random(10)

random(-R:float) is det

Binds R to a new random float in the open interval (0.0,1.0).

See also: - setrand/1, getrand/1 may be used to fetch/set the state.; - In SWI-Prolog, random/1 is implemented by the function random_float/0.

random_between(+L:int, +U:int, -R:int) is semidet

Binds R to a random integer in [L,U] (i.e., including both L and U). Fails silently if U<L.

random(+L:int, +U:int, -R:int) is det

random(+L:float, +U:float, -R:float) is det

Generate a random integer or float in a range. If L and U are both integers, R is a random integer in the half open interval [L,U). If L and U are both floats, R is a float in the open interval (L,U).

deprecated: - Please use random/1 for generating a random float and random_between/3 for generating a random integer. Note that random_between/3 includes the upper bound, while this predicate excludes it.

setrand(+State) is det

getrand(-State) is det

Query/set the state of the random generator. This is intended for restarting the generator at a known state only. The predicate setrand/1 accepts an opaque term returned by getrand/1. This term may be asserted, written and read. The application may not make other assumptions about this term.

For compatibility reasons with older versions of this library, setrand/1 also accepts a term rand(A,B,C), where A, B and C are integers in the range 1..30,000. This argument is used to seed the random generator. Deprecated.

Errors: - existence_error(random_state, _) is raised if the underlying infrastructure cannot fetch the random state. This is currently the case if SWI-Prolog is not compiled with the GMP library.
See also: - set_random/1 and random_property/1 provide the SWI-Prolog native implementation.

maybe is semidet

Succeed/fail with equal probability (variant of maybe/1).

maybe(+P) is semidet

Succeed with probability P, fail with probability 1-P

maybe(+K, +N) is semidet

Succeed with probability K/N (variant of maybe/1)

random_perm2(?A, ?B, ?X, ?Y) is semidet

Does X=A,Y=B or X=B,Y=A with equal probability.

random_member(-X, +List:list) is semidet

X is a random member of List. Equivalent to random_between(1, |List|), followed by nth1/3. Fails of List is the empty list.

Compatibility: - Quintus and SICStus libraries.

random_select(-X, +List, -Rest) is semidet

random_select(+X, -List, +Rest) is det

Randomly select or insert an element. Either List or Rest must be a list. Fails if List is the empty list.

Compatibility: - Quintus and SICStus libraries.

randset(+K:int, +N:int, -S:list(int)) is det

S is a sorted list of K unique random integers in the range 1..N. Implemented by enumerating 1..N and deciding whether or not the number should be part of the set. For example:

?- randset(5, 5, S).
S = [1, 2, 3, 4, 5].          (always)
?- randset(5, 20, S).
S = [2, 7, 10, 19, 20].

See also: - randseq/3.
bug: - Slow if N is large and K is small.

randseq(+K:int, +N:int, -List:list(int)) is det

S is a list of K unique random integers in the range 1..N. The order is random. Works as if defined by the following code.

randseq(K, N, List) :-
      randset(K, N, Set),
      random_permutation(Set, List).

See also: - randset/3.

random_permutation(+List, -Permutation) is det

random_permutation(-List, +Permutation) is det

Permutation is a random permutation of List. This is intended to process the elements of List in random order. The predicate is symmetric.

Errors: - instantiation_error, type_error(list, _).

partial_list(@Term) is semidet[private]

True if Term is a partial list.

setrand(+State) is det

getrand(-State) is det

Errors: - existence_error(random_state, _) is raised if the underlying infrastructure cannot fetch the random state. This is currently the case if SWI-Prolog is not compiled with the GMP library.
See also: - set_random/1 and random_property/1 provide the SWI-Prolog native implementation.